Inheriting an existing ESX environment, and am looking for some help.

[Cruicer]

I am getting ready to inherit a ESX environment consisting of about 10 or so ESX hosts and around 400+ VMs. Everything is SAN attached going thru a HP XP12000 to EVA disk. I am looking for some recommendations on where to start first to determine what kind of condition the overall environment is in, health checking if you will. I’ve been out of the VMware environment for a few years and am a little rusty on what I need to be looking for. Any quick hits to get a 1,000 foot view would be greatly helpful or any tools that may be out there that I can run against the systems would be nice too.

[juchestyle · Accepted solution]

Hey Man,

Welcome back, we missed you!

Check the simple things first, make sure vms have current tool sets. Check to see if the ESX hosts have current patches. Check to see if vms with 2 disks have both disks stored on the same lun or if they are different. See if vmotion works. Find out if users are happy about vmware, if they arn't things probably arn't working properly, if they are, that is a better sign. Find out if you have any documentation on the setup, if you do, things are probably well taken care of. Someone who documents usually does a decent job verse someone who doesn't, usually but not written in stone. Find out how much empty space you have on Luns, if you have none, you are in trouble, if you have some great. How big are your luns, do you use extents.

Start there man.

Respectfully,

Matthew

[RParker]

First thing you should do is ensure ALL ESX hosts are up to date, with firmware.

Then you can check the overall ESX cluster health by looking at the logs from the VC (Virtual Center) console. I assume you have VC setup?

Then since you took it over, I suggest you bring everything up to date, so you are working with the latest versions. Is HA (High Availability) enabled for your hosts?

right click the cluster name, and click 'edit settings' to see VMware DRS/HA configurations, and see if HA is enabled. That's a good place to start.

Then spot check each host to see if there are any obvious problems with logs and then do vdf -h for your VMFS volumes for orphaned VM's and check for duplicates, so you know what you are working with.

[davidjerwood]

Well, I would first start by drawing up a quick visio diagram of the physical and virtual layout of the esx infrastructure. Here is a useful stencil. http://www.vmguru.com/modules.php?name=Downloads&d_op=viewdownloaddetails&lid=4&ttitle=Visio_Stencil...

Then I would

1) Determine what versions and patch levels are running.

2) Document any Vmotion configuration

3) Document any HA configuration

4) Monitor performance on physical esx hosts.

That should give you a fair indication to what you are dealing with.

[Cruicer]

Good info...as far a disk space, what is the common recommedation for unused space. I know I don't want to take each LUN to 95% full, but is 80% full acceptable. I know that is a loaded questions and am only looking for some thing to gauge off of to see if I am in trouble.

[Cruicer]

I can't get to that stencil...if you have a blank copy can you e-mail it to me directly?

Thanks

Dan

[davidjerwood]

Sorry, try the following link.

http://www.vmguru.com/content/view/33/63/

[EllettIT]

I've used the "Veeam" reporter utility to get a quick overview of my ESX / VI3 environment, they have a monitoring software as well for VMware but I haven't used it. The trail of the reporter utility was fairly easy to use:

http://www.veeam.com/default.asp

[Dave_Mishchenko]

You'd want to stick closer to 80%. It'll depend on what you have on the LUN, but you'll want to leave room for snapshots and for the virtual machine memory swap files which will be equal to the size of memory that the VMs use minus any memory reservation.